Together AI
Verified May 27, 2026Together AI runs one of the largest open-weights inference catalogs, covering hundreds of models across the Llama, Qwen, DeepSeek, and Mistral families along with image and embedding endpoints. Pricing is per-token for chat and completion models, with separate per-second rates for dedicated endpoints and fine-tuned deployments.
The platform tends to keep pace with new model releases — typically within days of publication — and exposes both quantized and full-precision variants on the same API. Inference is served from US data centers with optional EU routing, and SOC 2 Type II is available on enterprise plans.
The trade-off versus aggregators is fewer routing options, but more transparent per-model SLAs and consistent latency across runs.
Calculate cost for your workload
Plug in your monthly tokens — get the actual bill on every provider serving each model.
Open calculator